Bearing fault diagnosis in 3 phase induction machine using current spectral subtraction with different wavelet transform techniques

Abstract

About sixty percent of the power in industries is consumed by induction machines, which implies induction machines are an integral part of industries. Even though these motors are stalwart and rugged in construction, they often experiences faults due to long time usage without maintenance. Bearing damage accounts 40% in the total faults and cause severe damage to the machine if unnoticed at nascent stage. So these faults should be continuously monitored for efficient operation, otherwise may cause severe damage to the machine. Conventional vibration monitoring is difficult due to requirement of high manpower and costly sensors. So motor current signature analysis (MCSA) is widely used for detection and localization of these faults. In this paper, the bearing faults are estimated by means of current frequency spectral subtraction using discrete wavelet transform. In addition to this, the current signature analysis after spectral subtraction is carried out using discrete wavelet transform (DWT), stationary wavelet transform (SWT) and wavelet packet decomposition (WPD) and a comparative analysis is presented to estimate fault severity using statistical parameters. The proposed model is assessed based on current signatures obtained from a 2.2 KW induction machine. The experimental results acknowledged the effectiveness of proposed method.